Cancer patients can develop metastatic disease years to decades after diagnosis. This pause can be explained by micrometastatic disseminated tumor cell (DTC) dormancy, a stage in cancer progression in which asymptomatic residual disease is present but clinically undetectable and resistant to conventional chemotherapy. The etiology of cancer dormancy and how to therapeutically address it have confounded scientists and clinicians. This is due in large part to the inherent difficulties in approaching this problem, including the rarity of DTCs and the surprising lack of genetic models of tumor dormancy. Although the Epithelial-Mesenchymal Transition (EMT) has been implicated in the initial stages of cancer metastasis, how it is maintained in late metastatic cells and whether it also regulates tumor dormancy remains unclear. We have recently detailed a mechanistic link between EMT and tumor dormancy in which Snail1 and Twist1, two key EMT-inducing transcription factors, cooperate with each other to initiate and maintain EMT, respectively, and in so doing Twist1 may contribute directly to tumor dormancy. Twist1 maintains EMT and perhaps tumor dormancy by inducing cell cycle arrest and a low ERK:p38 signaling ratio - two key features of tumor dormancy. The Snail1-Twist1 cooperation was observed in human breast cancer and the Twist1:Snail1 ratio in bone marrow DTCs was found to be highly predictive of distant relapses. Therefore, Twist1 represents a potential target to manipulate tumor dormancy for clinical benefits. To develop safe and effective Twist1-based applications, Twist1's role in EMT maintenance and tumor dormancy will need to be confirmed and characterized. To that end, we propose to develop a mouse genetic model of EMT maintenance and tumor dormancy and a cell-based system to determine the mechanism of Twist1-induced growth arrest: In specific aim 1, we take advantage of the MMTV-NeuNT mouse model of breast cancer which develops ErbB2- positive breast tumors de novo. We have shown that the Snail1-Twist1 temporal and spatial cooperation exists and dormant DTCs are present in this mouse model, establishing it as a bona fide genetic model of human breast cancer EMT and dormancy. To confirm that Twist1 is critical for maintaining tumor dormancy, we will determine the ability of Twist1- depleted tumor cells to maintain dormant DTCs, give rise to metastases, and confer treatment resistance as compared to control tumor cells. In specific aim 2, we have shown that the cytokine IL-6 is induced by Twist1 in late EMT and an important intermediary between Twist1 and p38. Using transient TGF?1-treated epithelial cells and MMTV-NeuNT mice, we will determine how Twist1 induces IL-6, whether IL-6 is critical for Twist1-dependent EMT maintenance and tumor dormancy in vivo, and whether other cytokines or regulatory proteins are involved in Twist1-associated autocrine and paracrine cytokine network. In specific aim 3, we will collect additional paired primary human breast tumors and BM DTCs to determine the predictive and prognostic power of the Twist1:Snail1 ratio in BM DTCs as compared to other biomarkers. These 3 aims will allow us to critically address a role for Twist1 in EMT maintenance and tumor dormancy. The result will hopefully provide new targets to control tumor dormancy and stimulate research into this significant problem.